Healthcare-associated infections (HAI) are a significant issue in healthcare facilities worldwide. Hand hygiene (HH) remains the most effective method for preventing the incidence of HAI in routine patient care. Past and current interventions focused on the overall improvement of HH compliance, but studies found that the amount of time required to achieve full HH compliance with the existing guidelines may not be practical. Improving HH compliance at critical moments during patient care may be more effective than improving HH compliance at all opportunities. However, there are little to no studies on healthcare workers’ (HCWs) behavior regarding HH during the patient care process. Secondary data analysis on a prospective dataset from the STAR-ICU trial was completed to identify HCWs’ behavior patterns regarding HH during the patient care process. Multiple logistic regression for transitions with random effects using repeated measures and transition modeling was used to identify possible associations between HH compliance and patient care tasks, the order of tasks, and workload. The models adjusted for the effects of HCW type, glove use, and isolation precautions. The study identified 28,826 task sequences and 42,349 HH opportunities. HCWs were slightly less likely to do HH before critical tasks compared with other tasks (OR: 0.97, 95% CI: 0.96-0.99), but more likely to do HH after contaminating tasks compared with other tasks (OR: 1.12, 95% CI: 1.10-1.13). HCWs are also more likely to move from task sequences that have a relatively lower risk to patients to task sequences that have a relatively higher risk to patients than vice versa (65.4% versus 34.7%). HCWs are also less likely to do HH after moving from tasks that have a relatively lower risk to patients to tasks that have a relatively higher risk to patients than vice versa (OR: 0.93, 95% CI:0.92-0.95). HCWs’ HH compliance rates decreased as the workload level increased (OR: 0.93, 95% CI: 0.89-0.98). Workload did not appear to affect HH compliance before critical tasks or after contaminating tasks and did not affect the order in which HCWs perform patient care tasks. Increase in workload was associated with an increase in the odds of critical tasks occurring (OR: 1.55, 95% CI: 1.45-1.65). In conclusion, HCWs are more likely to perform HH after contaminating tasks to prevent contaminating themselves and to reduce the risk of transmission in subsequent task sequences. However, they do not perform tasks in an order that minimizes risk to the patient; instead, it appears that they perform tasks as they come up in routine care. Furthermore, HH is not being performed at critical moments during patient care. Lastly, workload did not affect the order in which HCWs perform patient care tasks, suggesting that HCWs behavior patterns contribute significantly to how they care for patients and perform HH. Interventions targeting the order in which HCWs perform patient care tasks and improving HH compliance before critical tasks may be more effective than those designed to improve HH compliance at all HH opportunities for reducing HAI rates.